Computing resource service providers and other service providers often may utilize numerous servers, data stores and other resources to enable users to provision a variety of computing resources for use in its operations. In order to reduce the risk of data loss, these service providers may allocate these computing resources in different hardware groupings (e.g., data centers, rooms within data centers, server racks, etc.) to provide data redundancy and ensure availability of these computing resources in the event of a failure of any of these hardware groupings. However, allocation of these computing resources using different hardware groupings may present certain disadvantages. For instance, allocation of computing resources across multiple hardware groupings may increase the network latency for communications between these computing resources. Additionally, in some instances, allocation of these computing resources in different hardware groupings may be unnecessary, since unavailability of one computing resource may result in the unavailability of any associated computing resources, obviating the need for such redundancy. Thus, in some instances, allocation of computing resources within different hardware groupings may result in added expense and inefficiency.